A friend and I have a running disagreement on which type of natural disaster causes the most deaths and destruction in the U.S. and worldwide–hurricanes, tornadoes, earthquakes, wildfires, floods, etc. In a given year any one of them could cause massive destruction (like December’s tragic tsunami disaster), but how about on average, and how has this changed over time (for example, floods in 1900 vs. tornadoes now)? –Andy B., via e-mail

No offense, Andy, but you’ve got a pretty narrow idea of what constitutes a natural disaster–although I’ll grant you it’s one that’s widely shared. The World Almanac, for example, has entries for hurricanes, tornadoes, floods, earthquakes, and the like but omits famines and epidemics, which generally are of natural origin, kill more people, and alter history to a substantially greater extent. One may argue that a natural disaster is a brief, impersonal convulsion in which humans are mere roadkill, whereas famines and epidemics take place over a longer period of time and human involvement is more central. But this seems to me a silly distinction that serves chiefly to eliminate the most shocking cases. Since the topic is dismayingly large, we’ll take it in chunks:

Worst disasters. The closest thing I’ve discovered to a comprehensive disaster listing is EM-DAT, the Emergency Events Database maintained by the Centre for Research on the Epidemiology of Disasters (CRED) under UN auspices. EM-DAT lists 9,620 natural disasters since 1900, including famines and epidemics, but even so has conspicuous omissions, a matter I’ll return to. As it is, of the 20 deadliest disasters, eight are droughts, seven are epidemics, two are famines (presumably of the nondrought variety), and the remaining three are floods. The lowest death toll of any of these catastrophes was 393,000. For comparison, the deadliest tornado I ever heard of killed 695 people in Missouri, Illinois, and Indiana in 1925.

Pestilence. EM-DAT lists incidents by country and so falls down when it comes to reporting pandemics, which cross national borders and span years or decades. It has no comprehensive entries for the most destructive disease outbreaks of modern times, namely the influenza epidemic of 1918-’19, which killed 20 million by conservative estimate and possibly 50 million; and the AIDS epidemic, which has killed 20 million to date. Also not fully reflected, in part because it began too early, is the plague epidemic that started in China in 1894 and killed 10 million over the subsequent 20 years.

Famine. Famine is also problematic. Prior to 1900 famines were principally of natural origin, the prototypical example being the Irish potato famine of 1846-’49, which was caused by crop failure due to blight and led to a million deaths and mass emigration, emptying Ireland to a degree from which it has never recovered. In the 20th century, in contrast, many of the most devastating famines were primarily a consequence not of natural forces but of foolish or homicidal government policy. In the former category we have the Chinese famine of 1959-’62, during which 20 million died as a result of the ill-conceived but not fundamentally malevolent Great Leap Forward; in the latter, the Soviet famine of 1932-’34, in which Stalin’s campaign to destroy the prosperous kulak peasant class led to the deaths of 6 to 8 million in Ukraine and the north Caucasus. Of course, many “natural” famines still occurred, such as the Ethiopian famine of 1984-’85, in which nearly a million died following a drought.

Given the limitations of the data I hesitate to draw conclusions about disaster trends, averages, etc. A CRED chart purports to show that the average annual disaster death toll was high in the first half of the 20th century, much lower in the second half, and spiked up again after 2000.

Other catastrophes. Turning now to natural disasters in the almanac sense of the term, we note that (a) the most devastating incidents usually occur in Asia; (b) the big killers are earthquakes and floods (including tsunamis and tidal waves); and (c) mortality has not diminished noticeably over time. Some (I don’t say all) disasters since 1900 in which 100,000 or more died: 1911, Chang Jiang (Yangtze) basin, China, flood, 100,000; 1920, Gansu, China, earthquake, 200,000; 1923, Yokohama, Japan, earthquake, 143,000; 1927, Xining, China, earthquake, 200,000; 1931, Huang He (Yellow River) basin, China, flood, 3.7 million; 1935, China, flood, 142,000; 1939, China, flood, 500,000; 1948, Ashgabat, Turkmenistan, USSR, earthquake, 110,000; 1959, China, flood, 2 million; 1970, Bangladesh, cyclone-driven tidal wave, 300,000; 1971, Hanoi, North Vietnam, flood, 100,000; 1976, Tangshan, China, earthquake, 255,000 (official; estimated up to 655,000); 1991, Bangladesh, cyclone, 139,000; 2004, south and southeast Asia, tsunami, 280,931. For comparison, the deadliest short-term disaster in U.S. history I could turn up was the loss of 6,000 lives in the Galveston, Texas, hurricane of 1900. How lucky the country where the worst that happens to most disaster victims is they lose the house.

Art accompanying story in printed newspaper (not available in this archive): Illustration/Slug Signorino.